基于DT-CWT和S4VM的埋地排水管道堵塞故障识别研究  

A Blockage Dectection Method of Buried Drainage Pipeline Based on DT-CWT and S4VM

作　　者：李洋 冯早 黄国勇[1,2] 朱雪峰 LI Yang1＇2 ,FENG Zao1＇2, HUANG Guoyong1＇2 ,ZHU Xuefeng1＇2(1. School of Information Engineering ＆ Automation, Kunming University of Science and Technology, Kunming 650500, China;2. Mineral Pipeline Engineering Technology Research Center of Yunnan, Kunming 650500, China)

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500 [2]云南省矿物管道输送工程技术研究中心,云南昆明650500

出　　处：《电子科技》2018年第10期33-38,共6页Electronic Science and Technology

基　　金：国家自然科学基金(61563024;61663017);昆明理工大学引进人才科研启动基金(KKZ3201503015)

摘　　要：针对城市排水管道堵塞难以有效检测,以及检测过程中常规管道部件难以有效区分的问题,文中提出了基于双树复小波(DT-CWT)及安全半监督支持向量机(S4VM)的堵塞识别方法。该方法首先对管道中采集的声学信号进行双树复小波对分解,并对分解得到的各频段分量进行声压级变换;之后在有效分量上提取脉冲因子和平均声能量密度作为用于分类的特征向量;最后,将特征向量输入S4VM分类器进行训练与测试,从而达到对无标签的不同程度管道堵塞和三通件的数据聚类和故障识别的目的。Aiming at the problems that it was difficult to detect the obstruction of urban drainage pipes and todistinguish the conventional pipeline components, a method based on dual - tree complex wavelet transform （ DT -CWT） and sate semi- supervised support sector sachine （S4VM） was proposed in the paper. Firstly, the acousticsignals collected in the pipeline were decomposed by DT - CWT, followed by converting the acquired components intosound pressure level. Secondly, the pulse tactor and the average acoustic energy density were extracted from the ef-tective components as acoustical teatures, respectively. Finally, the teature vector was input to the S4VM classifierfor training and testing, so as to achieve the data clustering and fault identification for unlabeled varying degrees ofthe blocking as well as the pipe components.

关 键 词：导波 管道堵塞 双树复小波 半监督学习 S4VM 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]